Aug 18, 2008 (CIDRAP News) – Public health officials from Indonesia recently published an analysis of nearly all of the country's H5N1 avian influenza cases, revealing that death was more likely in those who received antiviral treatment late, were not part of a cluster, and lived in an urban area.
The study, published online Aug 15 in The Lancet, was authored by officials from Indonesia's Directorate General of Disease Control and Environmental Health, the country's health ministry, as well as authorities from laboratories and health organizations. It includes data from public health investigations and, when available, patients' clinical information.
The evaluation covered all confirmed human cases between Jun 22, 2005, when Indonesia recorded its first H5N1 infection, to Feb 1, 2008. Included were 127 patients, 103 (81%) of whom died.
The case-fatality rate (CFR) rose from 65% in 2005 to 86.8% in 2007. However, Indonesian officials say the rate has declined so far in 2008. According to a report that appeared Aug 15 on the Web site of the health ministry's avian influenza committee, the CFR from January through July was 84.2%, based on the 19 cases and 16 deaths recognized by the nation (as of this writing, the World Health Organization has recognized 18 cases with 15 deaths).
Only 2 of Indonesia's 127 infected patients were not hospitalized. One had a mild infection and received outpatient care, and one refused treatment and died at home.
Emerging patterns
A review of 108 clinical histories showed that symptoms during the first 2 days after onset were nonspecific in most cases. Thirty-two (30%) patients had fever and cough, and nine (8%) had fever and dyspnea.
Of the 125 patients who were hospitalized, 104 were diagnosed with pneumonia immediately or shortly after admission.
Eighty-eight (69%) of the case-patients were treated with oseltamivir, and the median time between symptom onset and treatment was 7 days (range 0 to 21). Patients who received the drug early were more likely to survive; those starting treatment more than 5 days after onset were more likely to die.
The authors report there were 11 case clusters that involved 28 patients. Infected patients who were not part of clusters were more likely to die, but researchers did not find any differences between cluster patients and noncluster patients in terms of when they presented to a healthcare facility, whether they received oseltamivir, or how soon they received the drug.
Patients with secondary cases were more likely to survive than primary case-patients, and they received antiviral treatment about 3 days earlier than primary case-patients. The investigators acknowledge that secondary cases may have involved other early interventions as well. They also report that patients who had indirect exposure to the virus were more likely to die.
Death and survival patterns among patients in clusters deserve further study, the authors state. Though close-knit families may be exposed to a common viral source, the role of genetic susceptibility and H5N1 virulence may also play important roles, they note. "Further studies should therefore be done on clusters to elucidate the definitive causes of reduced case fatality."
A need for new strategies
Most patients were hospitalized too late and received oseltamivir too late, the group says. "Training and equipping of all H5N1 referral hospitals across Indonesia, together with increasing the number of referral hospitals, is in progress to address this issue."
The authors emphasize that early identification is often difficult, but more information from agricultural officials about local poultry outbreaks could help healthcare workers increase their index of suspicion for H5N1 infections.
Other measures that could help reduce the country's CFR from H5N1 infections include rapid diagnostic tests for field use and better case-management training for healthcare workers, they write.
Experts call for more rigorous data
In a commentary accompanying the Lancet report, two British researchers say more uniform and complete data are needed to shed more clarity on trends emerging from Indonesia's cases—which account for a third of cases worldwide. The researchers are Sheila Bird, a biostatistician at Medical Research Council's biostatistics unit in Cambridge, and Jeremy Farrar, who directs the Oxford University clinical research unit at the Hospital for Tropical Diseases in Ho Chi Minh City, Vietnam.
"Basic clinico-epidemiological data are an essential adjunct to virological surveillance," they write. For example, exposures to the H5N1 should be clearly specified, including dates, they say, adding that lack of full data raises questions about such issues as whether the time frame between H5N1 disease onset and hospital admission has decreased over time.
"Indonesia, with the most extensive experience of human H5N1 patients, has a crucial clinical, epidemiological, and scientific role to play in the world's response to this potentially devastating infection," write Bird and Farrar.
The time to bolster national surveillance for human H5N1 cases is now, and disagreement over virus sample sharing should not hamper the flow of epidemiologic data, they add. "The world also needs to find a more equitable way to ensure that all share in the benefits of such important research. Indonesia could give the lead here."
Kandun IN, Tresnaningsih E, Purba WH, et al. Factors associated with case fatality of human H5N1 virus infections in Indonesia: a case series. Lancet 2008 (published online Aug 15 [Abstract]
Bird SM, Farrar J. Minimum dataset needed for confirmed human H5N1 cases. Lancet 2008 (published online Aug 15) [Extract]
See also:
Jul 17 CIDRAP News story "Reports examine high H5N1 death rate in Indonesia"
